Making Meaning Computable
Thematix is one of few organizations in the world with the ability to develop custom ontologies for use in automating business systems. We have developed ontologies for use in travel reservations systems, government intelligence automation and electronics and entertainment companies. Thematix is one of several parties behind the most significant ontology now in use in the financial services industry — the Financial Industry Business Ontology, or “FIBO”. If you’ve seen Google’s “Knowledge Graph,” you’ve seen ontologies in action.
But, first, what is an Ontology? Ontologies are rigorously structured, machine-readable concepts and terminology, relationships, rules and statements of fact about a domain. They are superb tools for content management for long-term retention and re-use. They enable software systems and to make inferences about data and come to relevant conclusions with regard to them, to search and retrieve disparate data over diverse sources, to perform on-the-fly association of information to suit particular, ad hoc purposes and queries – providing answers to questions rather than just ‘hits’ to searches.
Ontologies can be coupled with a reasoning system, allowing inferences to be made on the rules and facts. One example would be architecture of certain aspects of the ontology related to metrics and analysis such that both description logic and rule-based reasoning can be used over the ontologies, rules, and instance knowledge. Another would be to ensure that some aspects of the ontology can be used to drive high performance, highly distributed recommender services. “Smart search” and related recommendations presented via semantic technology are distinguished from database query languages and text search because the rules are separated from the query language and also separated from the schema of the facts. Facts and rules are expressed declaratively, not embedded in software or rigid database schema, and are thus much easier to implement and modify. Rules may be added and changed without affecting the facts. And, because rules define the terms in the search query, the queries are simple because the complex predicate expressions are in the rules, not the query.
Creating an ontology requires iterative explication and conceptualization, working closely with domain experts to articulate essential concepts that, until now, may have been only tacit parts of their expertise. Key system stakeholders, owners and resources for instance knowledge are identified.
The careful articulation of business functions and processes in terms of actors and system interactions – are vital to providing context, focus and project closure. We identify anticipated re-use and evolution pathways, as well as any critical standards and resources the system must interoperate with.
The ontologist and the domain expert to visualize and critique the model as it grows over time. The conceptual model simplifies technical and formal aspects of the ontology, and helps us to articulate architectural trade-offs, costs and benefits as well as elicit the kinds of questions that need to be answered by the ontology.
Thematix uses canonical ontological definitions, rather than re-invent them. Thus, namespace definitions, related metadata, units, measurements, standards, governance policies and other commonly used structures & vocabularies are imported and used as part of the ontology to the degree appropriate and practicable.
Thematix adheres to the principles of ontological analysis stated in IDEF5, which “provides a theoretically and empirically well-grounded method specifically designed to assist in creating, modifying, and maintaining ontologies,” and which when used leads to higher quality and lower costs.